Skip to content

Switching the Time Series Engine

Overview

This article describes how to switch the time series engine in Guance. This operation is applicable when switching from InfluxDB to GuanceDB.

Prerequisites

Deployment Information

Name Value
guance-insert Default Address guancedb-cluster-guance-select.middleware
guance-insert Default Port 8480
guance-select Default Address guancedb-cluster-guance-select.middleware
guance-select Default Port 8481

Switch Procedure

Backup Tables

CREATE TABLE `main_workspace_bak` LIKE `main_workspace`;


INSERT INTO  `main_workspace_bak` SELECT * FROM `main_workspace`;


CREATE TABLE `main_influx_db_bak` LIKE `main_influx_db`;


INSERT INTO  `main_influx_db_bak` SELECT * FROM `main_influx_db`;

Insert GuanceDB Information

  • Log in to the Guance df_core database and execute the following statement:
INSERT INTO `main_influx_instance` (`uuid`, `host`, `authorization`, `configJSON`,`dbcount`, `user`, `pwd`, `dbType`, `priority`, `status`, `creator`, `updator`, `createAt`, `deleteAt`, `updateAt`)
  VALUES
    ('276f8404252347d18e469f65f5a1abdd', 'http://0.0.0.0', '{\"admin\": {\"password\": \"*\", \"username\": \"*\"}}','{"read": {"host": "http://guancedb-cluster-guance-select.middleware:8481"}, "write": {"host": "http://guancedb-cluster-guance-insert.middleware:8480"}}', 0, '', '', 'guancedb', 50, 0, '', '', UNIX_TIMESTAMP(), -1, UNIX_TIMESTAMP());

Verification:

select count(1) from main_influx_instance where uuid='276f8404252347d18e469f65f5a1abdd';

+----------+
| count(1) |
+----------+
|        1 |
+----------+
1 row in set (0.00 sec)

If the query result is greater than or equal to 1, the execution is successful.

Replace InfluxDB with GuanceDB Information

  • Change the InfluxDB status
update main_influx_instance set status=3 where status=0 and dbType='influxdb';

Verification:

select count(1) from main_influx_instance where status=3 and dbType='influxdb';

+----------+
| count(1) |
+----------+
|        1 |
+----------+
1 row in set (0.00 sec)

If the query result is greater than or equal to 1, the execution is successful.

  • Batch update the time series database information for all workspaces
update main_influx_db set influxInstanceUUID=(select uuid from main_influx_instance where status=0 and dbType='guancedb' order by priority desc limit 1) , dbType='guancedb' where status=0;

Verification:

SELECT count(1)
FROM main_influx_db AS db
JOIN main_influx_instance AS instance ON db.influxInstanceUUID = instance.uuid
WHERE db.status = 0
  AND db.dbType = 'guancedb';

+----------+
| count(1) |
+----------+
|       26 |
+----------+
1 row in set (0.01 sec)

If the query result is greater than or equal to 1, the execution is successful.

  • Batch update the metric value in datastore of the main_workspace table
update main_workspace set datastore=JSON_SET(datastore, "$.metric", "guancedb") where status=0 and datastore ->> "$.metric" = "influxdb";

Verification:

SELECT count(1)
FROM main_workspace
WHERE status = 0
  AND datastore ->> "$.metric" = "guancedb";

+----------+
| count(1) |
+----------+
|       26 |
+----------+
1 row in set (0.00 sec)

If the query result is greater than or equal to 1, the execution is successful.

  • Update the DB ID
UPDATE df_core.main_influx_db 
SET df_core.main_influx_db.db =(
  SELECT df_core.main_workspace.id
  FROM df_core.main_workspace
  WHERE df_core.main_workspace.dbUUID = df_core.main_influx_db.UUID
);

Verification:

SELECT count(1)
FROM df_core.main_influx_db AS db
JOIN df_core.main_workspace AS ws ON ws.dbUUID = db.UUID;

+----------+
| count(1) |
+----------+
|       26 |
+----------+
1 row in set (0.00 sec)

If the query result is greater than or equal to 1, the execution is successful.

Restart All Kodo Services

kubectl delete pods --all -n forethought-kodo

Clear Redis

Choose one of the two methods:

  • Restart Redis
kubectl get po -n middleware | grep redis
kubectl delete po -n middleware <redis pod name>
  • Execute the cache clearing command
# Delete all "dbinfo:*" cache records, execute the following batch delete command
redis-cli -h <host> -p <port> -n 0 -a <password> keys dbinfo:* | xargs -r -t -n1 redis-cli -h <host> -p <port> -n 0 -a <password> del

# Delete all "wkspInfo:*" cache records, execute the following batch delete command
redis-cli -h <host> -p <port> -n 0 -a <password> keys wkspInfo:* | xargs -r -t -n1 redis-cli -h <host> -p <port> -n 0 -a <password> del

# Delete all "tkn_info:*" cache records, execute the following batch delete command
redis-cli -h <host> -p <port> -n 0 -a <password> keys tkn_info:* | xargs -r -t -n1 redis-cli -h <host> -p <port> -n 0 -a <password> del

Feedback

Is this page helpful?